# OP2 — full text

> The complete written content of https://op2.par2labs.com, as Markdown, for AI assistants and answer engines. OP2 is a product of PAR2 LABS (https://par2labs.com). The one-line-per-page index is at https://op2.par2labs.com/llms.txt. Generated from the same records the pages render, on 2026-09-22. Images are described, not linked.

## In short

**The writers' room, rebuilt as software.** Twenty-one specialist agents develop a feature screenplay in twelve stages, and grade each one before the next may build on it.

#### What it solves

- A chatbot writes the whole film in one pass, with nothing checking it. → Specialists work in dependency order, and every pass is graded before the next builds on it.
- A large cast collapses into one voice. → Each character is a persistent voice with its own wound and speech, audited for sameness.
- A finished draft, and no idea what to fix first. → A script audit: coverage, a structure map, scene-by-scene verdicts and a ranked fix list.

#### Why it is different

- **21 Specialist agents** — Each owns one craft and reads every decision made before it.
- **3 Quality gates a stage** — Diagnose, extend, grade. Weak work is rewritten, not passed on.
- **1 Project, not a chat log** — Research board, cast bible, beat board and script editor, all editable.
- **4-act Built for Indian cinema too** — CBFC ratings, and Telugu written natively rather than translated.

Built for: Screenwriters with a premise, Writers with a finished draft, Indian-language productions, Anyone tired of chat transcripts.

## The product

### OP2

https://op2.par2labs.com/

AI Screenwriting Studio · Early Access · v2.0

An AI screenwriting studio that develops a feature screenplay in dependency order — cast, world, beats, scenes, dialogue — and grades every stage before the next is allowed to build on it.

The writers' room, rebuilt as software.

OP2 — Oscar Papa 2 — is a screenwriting studio, not a chatbot. Twenty-one specialist agents develop an industry-formatted feature draft in twelve dependency-ordered stages, from the story's creative contract through world and cast, conflict and emotional design to beats, scenes and dialogue, and every pass is graded before the next stage may build on it. The result is a project rather than a chat log: research board, cast bible, beat board, script editor and score dashboard, all editable. It also audits scripts you have already written.

Start writing: https://scriptlabapp.com — It runs at scriptlabapp.com, and every new account starts with 1,500 free credits — enough for a real generation, not a locked demo.

*Screens — From a premise to an audited draft.*

#### Proof

**The room, counted.**
  Stats:
  - **21** — Specialist agents, each owning one craft
  - **12** — Pipeline stages, built in dependency order
  - **3** — Quality gates at every stage
  - **8** — Genres, each judged on its own craft axes

#### Features

- **Twenty-one specialists** — Each stage is owned by an agent built for that craft, reading every decision made before it.
- **Quality gates** — Every pass is scored, checked for substance and graded before the next stage may build on it.
- **Characters that push back** — Persistent voices with their own wound and speech write their own lines, audited for sameness.
- **Emotional architecture** — The film's emotional curve is designed as its own layer, before a single beat is placed.
- **A project, not a chat log** — Research board, cast bible, beat board, script editor and score dashboard, each an editable tab.
- **Script audit** — Coverage, a structure map, scene-by-scene verdicts and a ranked list of what to fix first.

#### What it is for

- You have a premise and want a structured, formatted first draft to own and rewrite, not a chat transcript.
- A large cast keeps collapsing into one voice, and you want each character written from their own wound, beliefs and knowledge.
- You have a finished draft and want coverage, a structure map and a ranked fix list before another month of rewriting.
- You are writing for an Indian audience and need the 4-act structure, a CBFC rating and Telugu written natively rather than translated.
- You want to start a long run in the evening and find it finished in the morning, without babysitting it.
- You want an existing script rebuilt as a project — cast, locations, relationships, beats and scenes — to develop the next draft in.

#### Specifications

- **Agents**: 21 specialists, not one generalist
- **Pipeline**: 12 stages, dependency-ordered with quality gates · Story DNA to screenplay
- **Quality gates**: 3 per stage · diagnose, extend, grade against the hit-screenplay bible
- **Genres**: 8, each with its own craft rules · 17 adaptations
- **Structures**: 3-act · hero's journey · 5-act · Indian 4-act
- **Ratings**: CBFC · MPAA certificate ratings
- **Languages**: English · Telugu transcreation · more Indian languages on the roadmap
- **Models**: Story-tuned LLMs · model chosen per agent
- **Modes**: Feature · series · short
- **Starting points**: Guided setup · a one-paragraph description · an existing script (PDF, Word, text, Fountain)
- **Formats**: Screenplay PDF · project PDF · Final Draft (.fdx) · Fountain · plain text · project bundle
- **Export**: One-click production draft · storyboard sheets · shot list (CSV or print)
- **Collaboration**: Real-time co-write · inline comments
- **Runs**: Durable jobs · retry and resume · live per-stage cost

#### Features — A real project, not a chat log.

https://op2.par2labs.com/features

The seven things that make it a studio rather than a chatbot, the tabs they live in, every screen of the running app, and a straight comparison with asking a chat model, dimension by dimension.

**Every capability, grouped by the job.**

**Every screen, one at a time.**

**Against one prompt and one model.**

| Dimension | One prompt, one model | OP2 |
| --- | --- | --- |
| Shape of the work | One prompt, one long answer. | A dependency-ordered pipeline of specialist stages. |
| When the output is weak | You re-prompt and get a different roll of the dice. | A critic diagnoses it and directs a targeted rewrite, keeping what worked. |
| Consistency at 110 pages | Degrades with distance from the prompt. | Every stage reads every prior decision as a binding constraint. |
| Character voice | One narrator wearing everyone's hat. | Persistent character voices, audited for sameness. |
| Emotional design | Implicit and accidental. | Designed as its own layer, before the beats exist. |
| What you receive | Text in a chat window. | A project with a research board, cast bible, beat board, score dashboard and storyboard. |
| Long jobs | Times out and needs babysitting. | Durable, resumable and self-healing. |
| Non-English | Literal translation that reads like translated English. | Transcreation written natively inside the locked structure. |

*The comparison is with a general chat model given one prompt, not with other screenwriting software. It cuts the other way on time and cost: a chat model answers in seconds, while a feature through OP2 takes hours and spends credits at the real AI cost of every stage.*

#### Audit — Already have a script? Get it read.

https://op2.par2labs.com/audit

Upload a finished draft and OP2 returns a professional-grade read: coverage, a structure map, a marketability call, scene-by-scene verdicts and a ranked list of what to fix first — judged purely on the page, with no reputation bias.

**What comes back.**
  Items:
  - **Professional coverage** — A logline and synopsis, plus an industry PASS / CONSIDER / RECOMMEND call for both the script and the writer — judged purely on the page, with no reputation bias.
  - **Positional structure map** — Where your inciting incident, midpoint and climax actually land, against where convention says an audience needs them.
  - **Marketability read** — Audience, budget tier, castability and comparable titles — the questions a development executive asks first.
  - **Scene-by-scene verdicts** — Mega mode rates every scene with what works, what is missing and how to fix it, plus a cross-scene pass that catches unpaid setups and flat runs.
  - **Per-character arcs** — Each character's want, cost and trajectory broken out, and where an arc stops moving.
  - **Ranked next-draft list** — One prioritised, scene-anchored fix list, so you know what to repair first before another month of rewriting.

**How it reads a script.**
  Items:
  - **PDF or pasted text** — Upload a PDF or paste the text of any screenplay.
  - **A score out of a hundred** — Broken down into scene success, craft and emotional impact, with exactly how the score was reached.
  - **The verdict** — A plain verdict on the film, the share of scenes that work, and what the next draft must fix.
  - **Characters, twice** — Each character scored as a concept and as they actually play on the page.
  - **Kept on record** — Every audit stays on record, and audited scripts can be compared by score, scene count and emotional journey.
  - **On the page only** — The read is of the script in front of it. Who wrote it carries no weight.

**Rewrite by element.**
  Items:
  - **One change, described once** — Describe a change to a character or a relationship in a sentence.
  - **Followed through the script** — It finds every scene where the change belongs, and the rewrite follows it through the screenplay.
  - **Handed on** — The audit exports as a PDF.

**Two doors into an existing script.**

| Door | What happens |
| --- | --- |
| Audit it | A scene-by-scene craft review with a PASS / CONSIDER / RECOMMEND call, a marketability read, a structure map and a ranked fix list. The fastest way to know where a draft actually stands. |
| Import it as a project | The studio rebuilds your cast, locations, relationships, beats and scenes as a full project, so you can develop the next draft inside the same governed pipeline that writes from scratch. |

*An import takes a PDF, Word file, plain text or Fountain file.*

#### Architecture — One stage at a time, in the order a story needs.

https://op2.par2labs.com/architecture

Each of twelve stages is owned by an agent built for that craft. It reads every decision made before it, does its work, and is graded before the next stage may build on it — so quality compounds instead of drifting.

**How it is put together.**

**The twelve stages.**
  Items:
  - **01 · Story DNA** — The creative contract: the thematic spine, the moral question and both sides of it, the four-corner opposition, and the emotional promise the film makes.
  - **02 · Research** — Grounded specifics instead of movie-generic ones. Sourced material is marked sourced; invented material is marked hypothesis.
  - **03 · World** — Locations with a visual and sound identity, symbolic weight and the rules of this world — not set dressing, but the world each scene is shot in.
  - **04 · Cast** — A cast bible: role, archetype, what each character knows and hides, motives, emotional wound, contradictions and how they speak.
  - **05 · Embodiment** — The cast stops being descriptions and becomes voices — persistent perspectives that debate their own profiles and write their own lines.
  - **06 · Conflict** — What collides and why, the stakes, and the irreversible moral cliff at the centre. Merely difficult choices are rejected here.
  - **07 · Structure** — The runtime carved into sections, each with act placement, structural function, timecode, emotional quotient and pacing.
  - **08 · Emotional design** — Elevation, humour and confrontation designed as their own layer, and one synthesised emotional curve for the whole film.
  - **09 · Beats** — Several complete beat boards, each beat carrying act, role, characters, location, emotional target and intensity. The strongest board advances.
  - **10 · Scenes** — Every beat expanded into scene cards: who wants what, what is in the way, what it costs, and the subtext underneath.
  - **11 · Dialogue** — Lines written per character, in that character's voice, using the real places and terms the research surfaced.
  - **12 · Screenplay** — Compiled to industry-standard format and timed to your runtime on the page-a-minute standard.

**Three gates at every stage.**
  Items:
  - **01 · Diagnose — is it well made?** — A critic scores the pass against that stage's craft criteria and returns a verdict on what to do next, not just a number.
  - **02 · Extend — is it enough to build on?** — Separately: is the pass substantive enough for the next stage to do its job? If it is thin, it is extended exactly where it is shallow.
  - **03 · Grade — does it meet the house standard?** — Graded against the hit-screenplay bible, the encoded standard of what working screenplays do that failing ones don't.
  - **The hit-screenplay bible** — PAR2 LABS' standard covers premise, structural turns, character want and cost, scene craft, emotional escalation and climax convergence. It adapts per genre. What it evaluates is published; how it is weighted is not.

**The specialists.**
  Items:
  - **One agent per craft** — Each stage is owned by an agent built for it, reading everything decided before it — a dependency-ordered team rather than one generalist answering a single prompt.
  - **Parallel where it can be** — Work runs in parallel where the story allows and in strict order where it does not.
  - **Grouped by role** — The Agent Garage lists every agent by role: orchestrator, generators, knowledge, scoring and audit.
  - **Story-tuned models** — The agents run on story-tuned language models, and the model each one runs on can be chosen in the Agent Garage, alongside the tools and the story bible the agents share.
  - **Nothing hidden** — The pipeline screen shows every stage as it runs, with how long it took, the tokens it used and what it cost.
  - **Rerun when the idea changes** — Change the idea and run the pipeline again.

**Three ways to start.**
  Items:
  - **Guided setup** — Structured choices: genre, tone, theme, era, region, runtime, certificate rating, audience, structure convention and narration pace.
  - **Describe your film** — One paragraph in plain language. The studio reads it, fills in the setup faithfully, and shows a logline, plot line and synopsis for you to approve.
  - **Upload an existing script** — A PDF, Word file, plain text or Fountain file, rebuilt as a full project: characters, locations, relationships, beats and scenes.

**The specification.**

#### Pricing — Pay for what you actually run.

https://op2.par2labs.com/pricing

OP2 runs on credits, where one credit is one cent of real AI cost. Every account starts with 1,500 free credits on signup, and pack prices are computed live, so the authoritative figure is the one shown at checkout in the app.

**What it costs.**

**How credits work.**
  Items:
  - **One credit, one cent** — A published conversion. You pay for the compute a generation actually consumes, not an opaque markup or a seat you have to guess the value of.
  - **Plan a run before you commit** — A planning calculator prices your exact configuration against your balance, and deliberately over-provisions so a funded run does not stall halfway for want of a few credits.
  - **Runs do not die on you** — If credits run out mid-run, the pipeline pauses and picks up where it stopped the moment credits land. Work in progress is not discarded.

**Questions we actually get.**
  Items:
  - **What do I get for free?** — Every account starts with 1,500 free credits on signup — enough for a real generation, not a locked demo — so you can run something meaningful before you buy a pack.
  - **Why credits instead of a flat subscription?** — Because the work varies enormously in cost. A feature-length generation with full critique passes costs meaningfully more than a short audit, and a flat fee would either overcharge the light user or undercharge the heavy one. Credits track real cost, one credit to one cent, with the conversion published.
  - **Can I predict what a feature will cost?** — Yes. The planning calculator prices your exact configuration — genre, runtime, structure, options — before you press generate, and deliberately over-estimates so a funded run does not stall midway. You commit knowing the number rather than guessing at it.
  - **Where do I see current pack prices?** — In the app. Pack prices are computed live and move with exchange rates, so the authoritative figure is always the one shown at checkout. That is why this page lists credit amounts rather than fixed prices.
  - **How is this different from just asking ChatGPT to write a screenplay?** — A chat model makes every creative decision at once, in one pass, with nothing checking it. OP2 makes them in order, with specialists, and grades each one before the next is allowed to build on it. The difference shows up in act two, in whether your characters sound like different people, and in whether the ending is earned or merely arrives.
  - **Does it write the whole screenplay, or do I still have to write?** — It writes a complete, formatted draft, with the beat board, cast bible, world, emotional arc and score report behind it. It is meant as a fast, structured first draft that you then own: every stage is editable, and any scene, beat or character can be regenerated without disturbing the rest.
  - **How long does a feature take?** — Hours, not minutes. A governed pipeline with real critique passes is not a ninety-second trick, and that is the point. You start it and walk away; it survives restarts and finishes on its own, with live progress and per-stage cost while it runs.
  - **Will it sound like AI?** — That is the failure the architecture is built against. Characters are embodied as persistent voices with their own wounds, beliefs, vocabulary and knowledge boundaries; they write their own dialogue, and a distinctiveness check flags any two who start to sound alike. Dialogue also draws on the project's own research rather than defaulting to generic movie English.
  - **Can I use it on a script I already wrote?** — Two ways. Audit it, for a scene-by-scene craft review with a PASS / CONSIDER / RECOMMEND call, a marketability read, a structure map and a ranked fix list. Or import it as a project, and the studio rebuilds your cast, locations, relationships, beats and scenes so you can develop the next draft inside it.
  - **What is the hit-screenplay bible?** — PAR2 LABS' encoded standard of what working screenplays do that failing ones don't — premise, structural turns, character want and cost, scene craft, emotional escalation, climax convergence. Every stage of every run is graded against it, and it adapts per genre, so a comedy is not judged on a thriller's axes. We publish what it evaluates, not how it is weighted.
  - **Which languages does it support?** — The studio writes in English and transcreates into Telugu, writing natively inside your locked structure rather than translating word by word, with a consistency pass and extra care on the tonally load-bearing scenes. Export in English, Telugu, or both in one document. Additional Indian languages are on the roadmap and not yet available.
  - **What formats can I get my script out in?** — PDF in industry screenplay format, a designed multi-section project PDF, Fountain, plain text, a portable project bundle, storyboard sheets, and a shot list as CSV or for print.
  - **What happens if it fails halfway?** — It is built not to. Work runs as durable jobs that retry, resume from where they stopped, and degrade a weak stage rather than killing the run, and a server restart does not lose the generation. If credits run out, the run pauses and picks up again when they land. Otherwise a run ends when it finishes or when you cancel it.
  - **Can I control the structure, or does it impose one?** — You choose: 3-act, hero's journey, 5-act or the Indian 4-act model, plus runtime, narration pacing, tone, certificate rating and budget tier, which also scales the size of the cast. Every structural slot is visible and editable on the beat board.

## Who makes it

PAR2 LABS PVT LTD — Technology . AI . Strategy . Films.

We engineer world-class technology and AI products across intelligence, infrastructure, hardware, and story — designed to make sense, deployed to make a difference.

Studio: https://par2labs.com · Contact: ceo@par2labs.com · +91 90591 69238